Any ideas? Trying to decipher Syvecs myself!

[BlueBugEye]

Hi guys,

Need a bit of help working out what's up with my engine/ECU.

Basic spec, forged 2.5 with 2.0 AVCS heads, just been re-freshed with new HG and bearings which means it's running a very slightly lower compression ratio than when it was last mapped. Running from a a Syvecs S6 mapped by JGM.

On the year old fuel that was in the car there appeared to be a bit of a misfire under load (only driven gently about 500m) the car then failed its MOT with Lambda of 0.965 at 2500 (also failed on HC of 417) first test was Lambda 0.959 HC 1248 this is at the test station I usually use and it usually passes (Miltek Sports Cat in centre section).

I've run the fuel through to get to some new Momentum in the system and got the following readings from the Syvecs:
Lambda at Idle = 0.930
Lambda at 2.5k = 0.99 - 1.00

At 4k rpm I had cylinder knock readings of up to 6 with cylinder 3 being consistently highest.

I also had a warning 'Engine coolant Temp 1, low after warm up'

The misfire does appear to have gone with the new fuel but not yet driven under load.

Now the questions!

Do these values look about right?
How can I use the Syvecs to identify a misfire and try to work out the cause (Fuel or spark and which cylinder)
What does the error code actually mean and what could be the cause?

I'm no expert so keep the replies simple!

Thanks for your help.

[pat]

A few thoughts....

Lambda 0.93 at idle suggests closed loop lambda control at idle is disabled. It is normally a pain to tune CLL to work OK at idle and higher up so quite often the simplest option is to just turn it off. It is possible, with some of the later additions to the firmware, to tune it successfully for full time closed loop though.

Lambda 0.99 - 1.00 at 2.5k suggests that the closed loop IS working there, you can see what it is doing by looking at the variable fuelMltCll1. In any event there appears to be a mismatch between the calibration of your sensor and that of the test station. I would imagine the test station is right since they are using a 4 gas analyser and you are not Time to tweak the Lambda sensor linearisation under Sensors->Lambda Sensors->Lambda Sensor 1->Linearisation. Specifically, where it is now reading 1.0 it should be reading 0.963. Adjust the rest of the curve to suit. To find the true lambda=1 point you can use the rear O2 sensor if it is still fitted. If you bring up a gauge for the input that is connected to the OEM sensor (presumably an11V based on AVCS engine, presumably MY01-05) you should be able to see the point that it tips past 0.45V (lambda 1). If your wideband sensor also says 1.0 at the same point then it's good to go.

Finding the origin of the misfire can be aided by setting the minimum throttle for cylinder trim corrections to (say) 1% in Breakpoints and Thresholds -> Correction Breakpoints. That will allow you to tweak a cylinder at a time even at idle. In Run Mode Fuelling -> Corrections -> Cylinder Trim Multiplier you can then multiply the fuel delivered to each cylinder - mulitplying by zero should stop fuel from getting to that cylinder. Get the engine to a point where there is a miss, then shut down a cylinder at a time - the one(s) that don't make much difference are the culprits - the one(s) that make a bigger difference are working fine. Remember to restore to 1.00 (or whatever individual trim was there to start with).

The error code just means you let the car warm up without driving it. There is some plausibility testing of sensor readings to try to identify failed sensors. Basically what happened is that it took so long for the engine to warm up that the sensor was flagged as faulty - ie "it's simply not possible for it to take this long, therefore the sensor must be wrong!". The plausibility is based on the premise that you start the engine because you mean to drive the car, and under those conditions it is indeed right that it cannot possibly take that long, But if you let the car idle then it can take a long time, especially in cold weather. You can fix this by either extending the warmup time allowed, or driving the car once you've started the engine.

Hope this helps,

Pat.

[west_minist]

Pat,

If you had to log for misfire, what would you log to capture some sort of view, since there isnt a wb per header port?

[pat]

Logging for misfires is a little interesting because there's no specific misfire detection strategy. But you could get some idea, the crank wheel is pretty much a 36-6 so has teeth packed nicely at 10 degree increments after TDC. You could log trRpm at 1000Hz, that will give you the average across 30 degrees of crank rotation. After a cylinder fires you should see an increase in trRpm, you'll get one whether it fires or not, but the increase will be more when a pot lights up. All you need then is to know where the crank is at... you can most likely figure that out from logging cyl0XKnock at 1000Hz too. That value will change within 30 degrees of the knock window closing (normally in the region of 30 ATDC)..... so if trRpm rises reasonably quickly in the sample or two around a cyl0XKnock change then you know that cylinder X has lit up properly, whereas if it only rises slowly then cylinder X has probably not lit up - you should also see a fluctuation in the appearance AFR measured by the lambda sensor after the misfire, but the transport delay can be quite long due to the length of the header pipework.

Hope this helps,

Pat.

[west_minist]

Yes it help. Almost seems you will need a wb per exhaust port

Thank you.

[west_minist]

Can you send an updated list of the parameter or post online?

I'm not seeing what trRPM stands for but know it has to be related to raw rpm or related to the rpm pickup etc under Engine Configuration.

Please forgive me for so many questions. I am learning the Syvecs/Life Racing.

[pavlo]

trRpm is essentially a less filtered RPM reading, averaged typically every 30 degrees of crank rotation.

You could spend the next 6 months asking questions and still only be slightly closer to understanding it all. However get an ECU on a car and start using it and you will get up to speed very quickly.

As for the original question, your best bet is to see someone that knows what they're doing to solve your immediate issues, and then if you want to learn how to use the ECU yourself, either get stuck in and get occasional support here, or pay for some proper training.

[west_minist]

Yes. As I told Romain and as you may have been copied on some emails, that is the aim for me next year is to get a S6 as soon as possible.

[pavlo]

I'm not Pat btw!

[west_minist]

My apologies.

Happy New Year!